5 Possible Causes of Scarring Effects
5.2 Evidence: Career Risk vs. Market Discipline
The main result of the previous section is that hedge fund liquidations entail sig-nificant and persistent scarring effects, mainly on high-ranking managers. In itself this finding does not help us to discriminate between the “market discipline” and the
“career risk” hypotheses. One could argue that, given their decision-making power, high-ranking employees are subject to the greatest reputation loss. But they also are likely to be those with the most human capital at stake: they may have developed portfolio strategies, client relationships and work habits that cannot be easily trans-planted to a new job, possibly outside the hedge fund industry or even the finance industry altogether. Hence, they may stand to lose more than lower-ranking employ-ees. Thus the absence of a scarring effect for mid-level professionals is not sufficient evidence against the “career risk” hypothesis.
18Expression (12) also implies other intuitive comparative statics results regarding the optimal liquidation time t∗. The more severe the information asymmetry, the less tolerant investors are of persistently poor relative performance: the time to liquidation is decreasing in the severity of moral hazard (low productivity ∆ and high private cost cof managerial effort) or adverse selection (low quality of the manager poolλ). Intuitively, when information problems are worse, underperformance results in a sharper fall in the manager’s reputation, inducing investors to cut in the manager’s fees more deeply, and thus bringing forward the moment when he is no longer willing to exert effort.
Liquidation is also hastened if investors set a more demanding target rateα.
To discriminate between the two hypotheses, we explore whether the impact varies with the fund’s relative performance prior to liquidation. According to the model in Section 5.1, a fund liquidation should tarnish the reputation of its managers only if it follows poor relative performance, and even then only if such underperformance is sufficiently persistent as to be informative of the managers’ skill, rather than a chance unlucky draw. The “career risk” hypothesis, instead, predicts that a liquidation may affect fund managers’ subsequent careers in fortuitous circumstances as well, such as adverse sector or market-wide trends. Indeed if the parameterφ >0, also employees of funds liquidated in such circumstances may face a decline in earnings, as they are less productive in their new jobs. In the former scenario, liquidations reflect, at least partly, a re-assessment of a manager’s skill; in the second, they simply result from bad luck.
To test whether relative performance before liquidation affects post-liquidation career slowdowns, we estimate the following variant of equation (2):
yit=αi+λgt+γLpostit +δLpostit ×Pit−+it, (13) where Lpostit is a liquidation dummy equal to 1 in the five years after liquidation and 0 otherwise, and Pit− is a “poor performance” indicator, i.e. a dummy equal to 1 if the liquidation follows a period (alternatively, 1 year or 2 years) in which the fund’s average monthly return fell short of its benchmark. Equation (13) also includes individual fixed effects, αi, and separate time effects for the two subsamples of control employees, λgt, where g = 1 for the control individuals matched with the employees of under-performing liquidated funds and g = 2 for those matched with employees of well-performing funds.
The coefficient γ measures the effect on career outcomes when liquidation is pre-ceded by normal relative performance; δ captures the incremental effect of poor per-formance. A negative estimate of γ in equation (13) would imply thatφ > 0 in the model of Section 5.1, namely, that also fortuitous liquidations have scarring effects, while a zero estimate of γ indicates that such liquidations have no scarring effects, i.e., φ = 0 in the model. The estimate of δ instead measures the career slowdown due to reputation loss from liquidation, which, as suggested by the model, should be only present if the liquidation is preceded by underperformance for a sufficiently long period (the optimal waiting time t∗).
The resulting estimates are shown in columns 1, 2 and 3 of Table 7 for the job level, imputed compensation and job mobility. What varies between the two panels is
the time interval over which performance is measured. The rationale for measuring pre-liquidation performance over various time intervals is that its informativeness about the managers’ quality should be greater for longer periods, as high-frequency noise in returns gradually abates. In the top panel, performance is measured over the year before liquidation; in the bottom panel, over two years. The estimates of the coefficient γ are small and not significantly different from zero for job level and compensation (columns 1 and 2), regardless of the length of the period chosen;
hence, when there is no prior underperformance, liquidation has no scarring effect, i.e. the parameter φ = 0 in the model. By contrast, the estimates of the coefficient δ in these two regressions rise in absolute value between Panel A and Panel B, and in the latter they become significantly different from zero. This indicates that, as time-averaging increases the signal-to-noise ratio in data on pre-liquidation returns, the scarring effect of liquidation following underperformance are both greater and more precisely estimated, consistent with the market discipline hypothesis. When a liquidation is preceded by two years of underperformance, it triggers a job level drop of 0.35 notches larger than if the liquidation were preceded by normal performance, and an imputed compensation loss over $420,000 larger, which is 24.5% of their imputed compensation in the year before the liquidation. A similar estimate of the imputed compensation loss is obtained using time-varying imputed compensation:
as shown in Table B.1 in Appendix B, in this case the imputed compensation loss is about $358,000.
[Insert Table 7]
As these estimates condition on poor prior fund performance, one may have con-cerns about the reliability of imputed total compensation as a measure of the earnings capacity of fund managers around the liquidation date: insofar as their actual com-pensation is tied to their prior performance, managers of underperforming funds are likely to earn less than the variable compensation typically imputed to their job title, both before and after liquidation of their fund. Hence, to provide a lower bound for the change in compensation associated with the liquidation of underperforming funds, we re-estimate the specification of Table 7 using only the fixed component of imputed compensation. The corresponding results are reported in Column 1 of Table B.2: when liquidation is preceded by two years of underperformance, it triggers a
$44,000 drop in the fixed component of imputed compensation, i.e. 14.5% of their imputed fixed pay in the year before the liquidation (about $303,000 on average). So
the estimated loss is a sizable fraction of pre-liquidation pay even if one neglects the variable component of compensation.
By contrast, the effects of liquidation on job mobility do not appear to vary with pre-performance: column 3 indicates that liquidation is followed by an increase of 5 to 6 percentage points in the probability of switching to a new employer, with no significant difference when liquidation is preceded by underperformance. Even liquidations that imply no information regarding the affected employees, presumably induce some employees to switch to other companies for more suitable jobs. By the same token, the employees affected by liquidations preceded by poor performance (and by the associated reputation loss) have an equal probability of switching to a new employer, but suffer a career slowdown. This squares with the idea that the setback does not stem simply from the frictions associated with changing jobs.
To further corroborate the hypothesis that the scarring effects documented above are induced by reputation loss due to fund-specific underperformance rather than by low absolute returns, we estimate equation (13) on the sub-sample of funds with positive absolute returns in the 24 months prior to liquidation. The estimated coef-ficients, not reported for brevity, are very close to those reported in Table 7: even conditioning on positive absolute returns, liquidations preceded by persistently poor relative performance are associated with significant career setbacks. It is relative, not absolute, pre-liquidation performance that triggers scarring effects.
Interestingly, the labor market appears to penalize more severely the managers of funds that fall short of their benchmark in good times (namely, when the benchmark does well) than those that do so in bad times. A natural interpretation for this finding is that relative underperformance is a stronger signal of low managerial skill when it occurs in booming than in bear markets. We document this pattern by re-estimating equation (13) separately for the subsample where the benchmark features positive returns and for that where it features negative returns. Table 8 reports the coefficient of the liquidation dummy and that of its interaction with the relative prior (2-year) underperformance dummy for each of the two regressions.
[Insert Table 8]
When benchmark returns are positive (Panel A), the interaction coefficient is es-timated to be negative and significant, and larger in absolute value than in Table 7.
In contrast, when benchmark returns are negative (Panel B), the coefficient is not significantly different from zero. This finding contrasts with evidence from other in-dustries that underperforming top executives are less penalized when their industry
is doing well: Jenter and Kanaan (2015) document that “better peer group per-formance substantially reduces the probability that an underperformer is dismissed, which implies that many fewer underperformers are fired in good times than in bad times” (p. 2156). This difference suggests that the labor market for asset managers may be more effective in filtering out aggregate noise when evaluating individual performance than the boards of public companies.
Since the previous subsection shows that only high-ranking managers suffer sig-nificant career slowdowns after liquidations, it is worth investigating whether this happens only in the wake of persistent pre-liquidation underperformance. This pro-vides a sharper test of the thesis that the career slowdown arises from reputation loss among top executives. To implement this test, we re-estimate equation (13) separately for high- and mid-ranking employees. The results are reported in Table 9.
[Insert Table 9]
In our estimates, only high-ranking employees (those with level-5 or level-6 jobs two years before liquidation) whose funds were liquidated after underperforming their benchmarks for two years suffer a post-liquidation career slowdown. Panel A of Table 9 reports the estimates for high-ranking employees, Panel B reports those for mid-level employees (mid-level-3 or mid-level-4) two years before the liquidation. Columns 1, 2 and 3 show the results for the job level, compensation and mobility.
Liquidations after normal performance are not followed by significant change in either the job level or compensation of top employees, but those that come after persistent underperformance do have significant scarring effects. The interaction between liquidation and poor performance has a negative and significant coefficient in both the job level and compensation regressions: the job level drops by 0.44 notches and imputed compensation by $664,000 more than for top employees of funds that are liquidated in the wake of normal performance (i.e., 21.8% of their pre-liquidation imputed compensation). Also in this case, the result is robust to the use of time-varying imputed compensation: as shown in Table B.1 of Appendix B, using this variable the estimated loss would be about $563,000. The loss is sizable also if the specification of Table 9 is re-estimated using only the fixed component of imputed compensation: as shown in Column 2 of Table B.2, liquidation after two years of underperformance triggers approximately a $68,000 drop in this component of compensation, i.e. 15.8% of their pre-liquidation fixed pay (about $431,000 on average).
In our sample, liquidations after poor relative performance are the most common ones: 79% of the liquidated hedge funds performed worse than their benchmark in the previous two years.19 However, the job mobility of top employees increases after liquidation regardless of the fund’s previous performance: the probability of switching increases by 4 percentage points in the years following liquidations even of well-performing funds (though this coefficient is not precisely estimated).
To sum up, the scarring effects of liquidations preceded by poor performance are very large for high-ranking employees, but no significant effects are observable for mid-level employees. The evidence, then, is consistent with the idea that liquidations cause a career slowdown for managers who can be held responsible for their fund’s poor performance. This squares with the thesis that the scarring effects depend mostly on reputation loss, not the materialization of “career risk”. According to our model, therefore, such effects can be thought of as the source of “market discipline”, which serves as an incentive to fund managers over and above performance-based pay. Since fortuitous liquidations are estimated to have no scarring effects (φ = 0), our model suggests that the disciplining role of performance-driven liquidations is not diluted by career risk arising irrespective of performance.
6 Conclusions
We have found that, if finance professionals experience a great career acceleration upon entering the hedge fund industry, they also face significant setbacks and are more likely to switch to other employers following the liquidation of the fund they work for.
This “scarring effect” impinges only on high-ranking managers in the investment companies, and only on those whose funds significantly and persistently underperform their benchmarks. Top managers of funds wound up after two years of poor relative performance suffer job demotion and a sizable compensation loss. Instead, when it is preceded by normal performance, fund liquidation is not associated with career slowdown or significant compensation loss.
We interpret these findings using a model of asset managers’ careers featuring moral hazard and adverse selection, where the fund’s relative performance enables investors to gradually learn about managers’ skills, and both performance pay and
19Brown et al. (2001) also find that poor relative performance increases the probability of hedge fund termination.
the danger of liquidation play a disciplining role for managers. Liquidation may also have causes that are not performance-related, in which case they entail only career risk, and do not generate incentive effects – indeed, the frequency of such liquidations weakens the disciplining role of performance-related liquidations. In this framework, our empirical findings that performance-related liquidations are by far the most common, and that they are the only ones followed by substantial and persistent scarring effects, suggest that they play a strong disciplining role ex ante.
On the whole, our results reveal a new facet of market discipline in asset man-agement, operating via the managerial labor market. This labor market discipline is complementary to contractual incentives within the firm. The job market “stick”
may indeed be a corrective to the tendency to motivate asset managers by generous
“carrots”, i.e. performance-based remuneration that is far more sensitive to upside gain than to downside risk.
References
Adams, R. B., and T. Kirchmaier. 2016. Women on Boards in Finance and STEM Industries. American Economic Review 106:277–81.
Agarwal, V., N. D. Daniel, and N. Y. Naik. 2009. Role of Managerial Incentives and Discretion in Hedge Fund Performance. The Journal of Finance 64:2221–2256.
Axelson, U., and P. Bond. 2015. Wall Street Occupations. The Journal of Finance 70:1949–1996.
Baghai, R., R. Silva, V. Thell, and V. Vig. 2017. Talent in Distressed Firms: Inves-tigating the Labor Costs of Financial Distress. Working paper 2854858, SSRN.
Benmelech, E., and C. Frydman. 2015. Military CEOs. Journal of Financial Eco-nomics 117:43–59.
Bertrand, M., C. Goldin, and L. F. Katz. 2010. Dynamics of the Gender Gap for Young Professionals in the Financial and Corporate Sectors. American Economic Journal: Applied Economics 2:228–255.
Bertrand, M., and K. F. Hallock. 2001. The Gender Gap in Top Corporate Jobs.
Industrial and Labor Relations Review 55:3–21.
Brown, J., and D. A. Matsa. 2016. Boarding a Sinking Ship? An Investigation of Job Applications to Distressed Firms. The Journal of Finance 71:507–550.
Brown, S. J., W. N. Goetzmann, and J. Park. 2001. Careers and Survival: Compe-tition and Risk in the Hedge Fund and CTA Industry. The Journal of Finance 56:1869–1886.
Busse, J. A., A. Goyal, and S. Wahal. 2010. Performance and Persistence in Institu-tional Investment Management. The Journal of Finance 65:765–790.
Eckbo, B. E., K. S. Thorburn, and W. Wang. 2016. How Costly is Corporate Bankruptcy for the CEO? Journal of Financial Economics 121:210–229.
Edmans, A., X. Gabaix, and D. Jenter. 2017. Executive Compensation: A Survey of Theory and Evidence. ECGI Finance Working paper 514.
Gao, J., K. Kleiner, and J. Pacelli. 2017. Credit and Punishment: The Career Incentives of Wall Street Bankers. Kelley School of Business Research Paper 16-81.
Gibbons, R., and L. F. Katz. 1991. Layoffs and Lemons. Journal of Labor Economics 9:351–380.
Gibbons, R., and K. J. Murphy. 1990. Relative Performance Evaluation for Chief Executive Officers. Industrial and Labor Relations Review 43:30S–51S.
Goyal, A., and S. Wahal. 2008. The Selection and Termination of Investment Man-agement Firms by Plan Sponsors. The Journal of Finance 63:1805–1847.
Graham, J. R., C. R. Harvey, and M. Puri. 2013. Managerial attitudes and corporate actions. Journal of Financial Economics 109:103–121.
Graham, J. R., H. Kim, S. Li, and J. Qiu. 2017. Employee Costs of Corporate Bankruptcy. SSRN Working paper 2276753.
Griffin, J. M., S. Kruger, and G. Maturana. 2018. Do Labor Markets Discipline?
Evidence from RMBS Bankers. Working paper 2977741, SSRN.
Jenter, D., and F. Kanaan. 2015. CEO Turnover and Relative Performance Evalua-tion. The Journal of Finance 70:2155–2184.
Kaplan, S. N., M. M. Klebanov, and M. Sorensen. 2012. Which CEO Characteristics and Abilities Matter? The Journal of Finance 67:973–1007.
Malmendier, U., G. Tate, and J. Yan. 2011. Overconfidence and Early-Life Expe-riences: The Effect of Managerial Traits on Corporate Financial Policies. The Journal of Finance 66:1687–1733.
Murphy, K. 1999. Executive compensation. In O. Ashenfelter and D. Card (eds.), Handbook of Labor Economics, vol. 3, Part B, 1st ed., chap. 38, pp. 2485–2563.
Elsevier.
Oreopoulos, P., T. von Wachter, and A. Heisz. 2012. The Short- and Long-Term Career Effects of Graduating in a Recession. American Economic Journal: Applied Economics 4:1–29.
Oyer, P. 2008. The Making of an Investment Banker: Stock Market Shocks, Career Choice, and Lifetime Income. The Journal of Finance 63:2601–2628.
Pedregosa, F., G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blon-del, P. Prettenhofer, R. Weiss, V. Dubourg, et al. 2011. Scikit-learn: Machine learning in Python. Journal of Machine Learning Research 12:2825–2830.
Philippon, T., and A. Reshef. 2012. Wages and Human Capital in the U.S. Finance Industry: 1909-2006. The Quarterly Journal of Economics 127:1551–1609.
Schoar, A., and L. Zuo. 2017. Shaped by Booms and Busts: How the Economy Impacts CEO Careers and Management Styles. The Review of Financial Studies 30:1425–1456.
Table 1: Job Levels and Compensation
This table illustrates the two dimensions that characterize the employment positions of the indi-viduals in our sample: their job level, i.e. rank within the corporate hierarchy, and the typical compensation associated with that title and sector. Job levels are identified by first matching the job titles reported by individuals in their resumes with the Standard Occupational Classifica-tion (SOC) produced by the Bureau of Labor Statistics (BLS), and then grouping the SOC codes into six bins reflecting different degrees of decision-making power. To measure the average annual compensation associated in 2016 with each SOC code, for level 1-4 jobs we use the Occupational Employment Statistics (OES), allowing for differences in salary across the following six sectors: (i) asset management (AM), (ii) commercial banking and other lending institutions (CB); (iii) financial conglomerates, defined as institutions encompassing lending, insurance and/or asset management (CO); (iv) insurance (IN); (v) other finance, which consists mainly in financial consultancies and portfolio advisors (OF); and non-financial firms and institutions, including government, suprana-tional institutions and stock exchanges (NF). For levels 5 and 6, we use data on total compensation (including the variable component) drawn from the 10K forms filed with the SEC in 2015by com-panies belonging to the six sectors.
Job Average Examples of
Labors & Helpers, Service Workers 53,845 assistant, intern
Table 2: Descriptive Statistics
The table reports statistics on the characteristics of the individuals in our sample, based on data drawn from individual resumes available on a major professional networking website, together with information available from Bloomberg, Businessweek and companies websites. Education Level variables are indicators for the highest degree held. Subject variables designate the subject of the highest degree. The quality of highest degree is defined on the basis of QS Ranking, with three
The table reports statistics on the characteristics of the individuals in our sample, based on data drawn from individual resumes available on a major professional networking website, together with information available from Bloomberg, Businessweek and companies websites. Education Level variables are indicators for the highest degree held. Subject variables designate the subject of the highest degree. The quality of highest degree is defined on the basis of QS Ranking, with three